Reference : Biol. Bull., 141 : 247—260. (October, 1971)

REEF CORALS : AUTOTROPHS OR HETEROTROPHS?

THOMAS F. GOREAU,1 NORA I. GOREAU AND C. M. YONGE Discovery Bay Marine Laboratory, Department of Zoology, University of the West mndies, Mona, Kingston 7, Jamaica; and Department of Zoology, University of Edinburgh, Edinburgh, Scotland

Some recent studies 2 seem to indicate that the nutritional economy of reef corals is for all practical purposes to be considered autotrophic due to their zooxanthellae (Fig. 1). For example, Franzisket (1969a, 1970) claims to have demonstrated that some Hawaiian reef corals can achieve net growth in the total absence of particulate food, while Johannes and Coles ( 1969) state that the energy requirements of Bermudian reef corals are in some cases more than an order of magnitude greater than could be provided by the zooplankton which the investi gators were able to catch with a fine net. In spite of their supposedly autotrophic economy, the reef corals have not developed any of the behavioral and structural specializations for such a way of life. In this respect they differ fundamentally from Xenia hicksoni and Clavularia hanira (Octocorallia, Alcyonacea) (Gohar, 1940, 1948) and Zoanthus sociatus (Hexacorallia, Zoanthidea) (Von Holt and Von Holt, 1968a, b) , unrelated anthozoans which have independently evolved a more or less complete nutritional dependence upon their contained zooxanthellae. Available data is summarized in Table I. These species have never been observed to feed, and there is a more or less marked reduction of structures and functions associated with the usual predatory feeding habits in Cnidaria ; for example, they do not respond to any of the known tactile and chemical stimuli that trigger feeding behavior in related carnivorous species; they do not ingest particulate matter, and are unable to either digest or assimilate food artificially placed into their coelenteron by means of a canula (Goreau and Goreau, unpublished). The reef corals are, by contrast, superbly efficient and voracious carnivores that will accept practically any kind of particulate animal food (Yonge, 1930a, 1930b; Yonge and Nicholls 1930, 1931). Feeding occurs in several different ways, de pending on the species: in the majority, the food is swept into the coelenteron by means of ciliary currents, (sometimes involving reversal as in Fungia), while in some corals the tentacles convey the food directly to the mouth (Yonge, 1930a). Most species are also capable of extracoelenteric digestion of food matter outside the body by means of mesenterial filaments extruded through temporary openings (Fig. 2) at any place on the colony surface (Duerden, 1902; Matthai, 1918; Goreau, 1956). Reef corals obtain food via this ancillary route and also use the extruded filaments as weapons, primarily against other corals the tissues of which they may digest (Lang, 1969, 1970).

1 Thomas F. Goreau was Professor of Marine Sciences, University of the West Indies at Mona, and Professor of Biology, State University of New York at Stony Brook; Director of the Coral Reef Project and Director of the Marine Laboratory at Discovery Bay, Jamaica. He died unexpectedly in New York on April 22, 1970. 2 See also the paper by V. B. Pearse and L. Muscatine (dedicated to the late T. F. Goreau) on pages 350—363,and that by P. V. Fankboner on pages 222—234of this issue—Editor. 247 248 THOMAS F. GOREAU, NORA I. GOREAU AND C. M. YONGE

FIGuu.@ 1. Autoradiogran1 of carbon1' labelled Fungia scutaria. The coral was ex posed to ‘¿4C02in sunlight for ten minutes and washed in running sea water for thirty minutes before fixation. The area seen under (lark field illumination and focused on the plane of emul sion shows the concentration of silver grains over the carbon― labelled zooxanthellac of the gastrodermis. The scale represents 100 @. Specimens exposed in the (lark showed no fixation of the isotope. FIGURE 2. Mussa angulosa under severe starvation gradually loses contact with the skele ton (a) (b), eventually sinking to the bottom of the aquarium, still alive. When offered crab juice this free specimen extruded its mesenterial filaments (c) through the epithelium of the calicoblast. Some colonies lose their zooxanthellae, some keel) them, but in the most severe cases of starvation only the stoniodeum and a few filaments remain. Later only bits of mesenterial filaments curl about. However they all showed feeding responses when crab meat or amino acids were added. These filaments persist for a few (lays.

These diverse feeding mechanisms are supplemented by an exquisitely per ceptive chemotactic sense. In several species of Jamaican reef corals (Manicina areolata, Cladocora arbuscula, Eusnülia fastigiata, Isophvllia sinuosa., Mussa angulosa and Scolvm.ia lacera) we found some years ago that very low concen trations of amino acids such as glycine, alanine. phenvlalanine and lucine could trigger off typical feeding responses; i.e., opening and eversion of the stomodeum, swelling of the coenosarc, extension of tentacles and sometimes extrusion of mesenterial filaments. M. areolata responded in this manner to alanine and glvcine at concentrations as low as 10@ @i,whereas glucose, sucrose, glycerol and mannitol did not have any effect at high concentrations. Mariscal and Lenhoff (1968) ob TROPHIC CONDITIONS OF REEF CORALS 249

TABLE I Available data on nutritional adaptation and absorption in Scleraclinia, Alcyonacea and Zoanthidea

of of of re crab juice ‘¿Hleucine sponse to TaxonomyZooxanthellaeAbsorption and india ink into the C by crab meat or acidsNematocystsScieractinia: into filamentsUptakeepidermisUptakezooxantliellaeFeedingamino

Hermatypes: Fungia + Stylophora + +++ +++ ++ +++ +++ Ahermatype: Tubastrea 0 +++ +++ 0 +++ +++ Alcyonacea: Xenia + 0 ? +++ 0 0 Zoanthidea: Z. sociatus + ? ? +++ ? disordered P.caribbae + +++ ? + +++ ++ P. grandis ++++ ++++++ ?++ ++++ ++++++ ++

Other niorphological correllates with xanthellar symbiosis are: (1) Xenia: nonematocysts,reducedfilaments,noseptallobes, (2) Zoanthus: nematocysts,butthesearein a disorderedpositionandin placeswheretheydonogood; filaments reduced but lobesare very large. All stages of pycnosis, degeneration,frag mentation and extrusion of zooxanthellae were observed in the mesenterial lobes. Feeding reaction: (1) Corals: Dilationand extensionof stomodeum,inibibitionof water,sometimeserectionof tentacles or shooting of the mesenteries through mouth or body wall; (2) Xeniids: Rhythmic movement of tentacles of anthocodia; (3) Zoanthids:Zoanthussociatus:none, Palythoa caribbae : Dilation of mouth, strongly inward movement of water at ciliate groove, curling over of the lenkicular rinz, Palythoa grandis: same as P. caribbae.

served tilat concentrations as low as 10@ M proline resulted in feeding responses in C@'plzastrea ocellina., Pocillopora dai'nicornis and Fungia scutaria. Responses similar to tilose caused by amino acids are produced in corals by seawater ill which there had previously been zooplankton. We have often observed that corals will expand tinder llattlral conditions in apparent anticipation of plank ton: evidently this is due to tileir ability to sense the diffuse cloud of metabolites, including amino acids, that usually surrounds plankton swarms (Hellebust, 1965). It would indeed be surprising if, as Johannes and Coles (1969) have speculated, tilecoralshave retainedtllesecapabilitiesmerely to obtaintracenutrientssuch as phosphorus from their prey while the bulk of their nutrition comes from the zooxantheilae! Yet, there is no need for such a roundabout way to obtain phos phorus since reef corals, but not ahermatypes, in the light are known to take up inorganic phosphate from the mediunl, this being a function of the zooxantheilae, not the coral host (Yonge alld Nidllolls, 1931). The fully autotrophic xeniid alcyonaceans and Zoanthus are evidently able to obtain all their trace nutrients directly from tile seawater. As regards the possible need for organic phosphorus, Von Holt (1968) has shown that ill Zoanthus there is a transfer of nucleoside polyphospilate fronl algal symbiont to animal host. If the reef corals were truly 250 THOMAS F. GOREAU, NORA I. GOREAU AND C. M. YONGE as autotrophic as Franzisket and Johannes believe, the question arises why have they not evolved similar more direct mechanisms for obtaining critical nutrients directly from their symbionts?

OBSERVATIONS The boundary layer water and its relation to the trophic structure of the reef The evidence so far cited has not resolved the conflict between the apparent low productivity of tropical ocean surface waters (Fleming, 1954 ; Sargent and Austin, 1949, 1954) and the need for organic nutrients by the benthonic fauna in the reef ecosystem. This consists of the corals and a diverse assenlblage of filter, detritus, suspension and deposit feeders as well as predacious carnivores, the ma jority without zooxanthellae which might serve as ancillary food source. Recent reviews by Bakus ( 1969) and Stoddart ( 1969) have demonstrated how little quantitative information is available on the trophic cycles within the reef bio tope, largely because the pathways themselves are still largely unknown. Oceanic reef ecosystems appear on the whole to be autotrophic units operating at very high levels of productivity, turnover rate and efficiency (Odum and Odum, 1955; Kohn and Helfrich, 1957) whereas at least sonie smaller reefs off high islands may be non-autotrophic (Goreau, Torres, Mas and Ramos, 1960 ; Gorden and Kelley, 1962) . In view of the low trophic potential of the tropical oceanic waters, high localized productivity of reefs can only be achieved through coupled internal recycling systems that reduce external losses of free energy to a minimum and thus maintain the local nutrient levels at high steady state values. The existence of such internal cycles is reflected in the marked differences that may be observed between the outside ocean water and the water circulating within the reef which will be referred to here as the boundary layer water. Whereas the former is clear and deficient in plankton and other suspended matter, the latter is relatively turbid due to the much higher concentrations of suspended particulates, consisting of both inorganic and organic detritus stirred up by the turbulence, or added to the water by bentilonic biota. Near lligh islands, both particulate and dissolved nutrients in the sea are increased by run-off from the land. The boundary layer water also contains a relatively high concentration of zooplankters, swarms of which shelter and feed within the multitude of crevices and other microhabitats of the reef frame. This environment is extremely difficult to sample quantitatively, but can be readily observed by anyone diving on the reef. The depauperate and heavily cropped condition of the shallow reef zones de scribed by Bakus (1967, 1969) for some Pacific reefs, and by Johannes and Coles (1969) for Bermuda have been corroborated by the first author's own observa tions on parts of the Great Barrier Reef, Eniwetok, Saipan and the Red Sea, and is also observed in the shallow reefs of Jamaica and other Caribbean islands. However, conditions in the deeper parts of the outer reef slope vary considerably from extreme impoverishment as for example in Saipan or Eniwetok to a marked increase in species diversity, size and biomass of the macrobenthos, such as is observed in Jamaica (Goreau and Hartman, 1963; Goreau and Wells, 1967). Here the fore reef slope habitat is characterized by very large and diverse standing crops of corals, sponges, Gorgonacea, anemones, Antipatharia, and various algae such TROPHIC CONDITIONS OF REEF CORALS 251 as Halimeda (Goreau and Graham, 1967) ; the interstices of the reef frame contain an abundant fauna of Foraminifera, sponges (Hartman and Goreau, 1970) , hy drozoans, ahermatypic corals, worms, bivalves, brachiopods (Jackson, Goreau and Hartman, 1970) , bryozoans, echinoderms, tunicates and arthropods. Only the herinatypic corals, with the great majority of other coelenterates, contain zooxan theilae, the remainder do not. Above sixty meters the corals predominate, below this the sponges prevail although reef corals occur in diminishing amounts to at least one hundred meters. The boundary layer water is in continuous and dynamic exchange with the reef biota. We established this by releasing small clouds of India ink from syringes in various microhabitats of the Jamaican fore reef slope at depths of 50 to 60 meters where wave turbulence is low. We found that the India ink was cleared from the water within a few minutes, mostly by the sponges. It appears that a continuous downward flow of particulate matter moves from the boundary layer through the reef, recycling nutrients within the benthos. Quantitative measure ments of this exchange have now been carried out in situ by H. M. Reiswig (Biol ogy Department, Yale University) in Discovery Bay, Jamaica.

Suspended particulate matter in the reef as a possible food source for corals The particulate suspended organic matter, organic aggregates and dissolved organic substances circulating in the boundary water of the reef may be of crucial importance to the nutrition of the benthonic fauna, corals included. Marshall (1965)showedthattheamountoffinesuspendedorganicdetritalmatterinthe waters of Eniwetok Atoll was between one and two orders of magnitude greater than could be collected with the finest plankton nets. In Jamaica, the macroscopic organic particulates consist chiefly of comminuted vegetable matter, fragmented animal remains of diverse origin, faecal pellets, etc., but we have not yet investigated the much larger microscopic and submicroscopic fractions. Coral-browsing acanthu rid and scarid fish contribute large volumes of ground-up carbonates to the sus pended matter (Bardach, 1961). During periods of rough weather wave turbulence stirs up fine organic detritus, the leptopel, from the bottom sediment, and clouds of this material roll down over the reef communities of the seaward slope into deep water. At the same time, colloidal and dissolved organic matter are aggregated into larger particles at the surface of bubbles stirred up by the surf (Baylor and Sut cliffe, 1963; Riley, 1963). Mucus is secreted into the water in large amounts by ben. thonic animals in the reef, chiefly sponges, gorgonians, corals and molluscs (Mar. shall, 1965). Corals and alcyonarians continuously void large numbers of excess zooxanthellae in strings of mucus (Yonge and Nicholls, 1931). The gonadal products of sponges and echinoidsperiodicallyreach such high concentrationsas seriouslyto reduce underwater visibilityin the vicinityof the reef. The questionof whether any of these diverseorganic particulatesare available as food to the corals is still undecided. Part of the difficulty in relating reef corals to their potential food supply is a conceptual one. As the result of Yonge's studies (1940), the corals have been thought of principally as specialized planktivorous carnivores. However, we have numerous observations which seem to indicate that many of the reef corals are not restricted in their feeding to zooplankton since they also seem to feed on any organic particulates that happen to be carried into the 252 THOMAS F. GOREAU, NORA I. GOREAU AND C. M. YONGE coelenteron and from which nutriment may be extracted. In our experience, many reef corals are relativelyunspecialized detritus feeders (Fig. 3), capable of utilizing a wide range of organic matter and bacteria (R. A. Kinzie III. Department of Zoology. University of Georgia, personal communication). An example of this is the common Indo-Pacific reef coral Fungi@i scutaria (Goreau, Goreau, Yonge and Neumann, 1970) although it is not yet known what part of its total energy requirements are met from exogenous particulates other than zooplankton.

The uptake of dissolved organic matter b@'corals \Ve have not so far considered the possibility of direct utilization of dissolved or colloidal organic matter by scleractinian corals. No attempt will be made here to answer the question of whether there is enough dissolved organic matter of the right kind circulating within the reef to be a significant source of energy for corals; rather we wish to point out that the corals have highly developed struc tural and functional adaptations for the absorption of dissolved organic iliatter directly from the sea, and that they can take up compounds such as amino acids

FIGURE 3. Mcandrina nicandrites f. danae behaving as a detritus feeder is seen here sweeping the niud bottom with huge loops of mesenterial filaments extruded through the column wall after crab extract has been added to the media. These corals also showed feed ing reactions when offered alanine and glycine. @1@

TROPHIC CONDITIONS OF REEF CORALS 253

__ ‘¿

FIGURE 4. Autoradiogram of Fungia scutaria exposed to 3H leucine in sunlight for 1 hour, theii washed in running sea water. The radioactivity is restricted to the epidermal cells as shown by the (listribution of the dark silver grains in the overlying emulsion. Activity first appeare(l in the tall epidermal cells, then spread over the cellular tissue except for the zooxanthellae, mesoglea and niucus glands. Compare this with figures 6 and 7 for alkaline phosphatase and P.A.S. The (lark control shows the same pattern of activity phase contrast. from extrenielv dilute solution. There is also some Prelil1ii1iar@ evidence that certain Gorgonacea and Zoantliidea have similar abilities. In the stony corals, the absorption of dissolved organic matter takes Place mainly in the epidermis of the column wall, tentacles, oral disc and stomodaeum, i.e., the entire surface in (lirect contact with the external medium. Autoradiog raphy of the reef corals exposed to very low concentrations of tritiate(l DL leucine in sea water for one hour atid fixed at varying times after labelling show that the activity is initially fixed in the tall columnar cells of the epidermis, whereas much less is present in the gastrodermis, very little in the mesogloea and none in the zooxanthellae (Fig. 4). Twenty-four hours after labelling the activity is more uniformly spread throughout the cellular tissues, except for the zooxanthel lae. Epidermal uptake of amino acids by corals is independent of light intensity and absorption occurs even when the leucine concentration is below the threshold of chemotactic response of the test species, Fungia scutaria, about 3: l0@ @r. Ab sorption of glucose has since been observed by Stephens (1962). 254 THOMAS F. GOREAU, NORA I. GOREAU AND C. M. YONGE

CP

FIGURE 5. Electroniiiicrograph of the epidermal border of Astrangia danae tangentially cut. The epidermis consists of tall columnar cells with a finely granular cytoplasm the free surface of which bears a single flagellum (f) set into a shallow pit (pt) bordered by a circlet of nine to twelve microvilli. The microvilli are especially conspicuous on the collar processes (cp). At the base of the flagellum are shown the l)asal plate (b.p), basal corpuscle (b.c.) and rootlet fibre (r.f.). The latter has a periodicity of about 670 A. The arrangement of the microvilli suggests they are modified collars reminiscent of those of choanocytes. The surface membrane shows numerous and very variable cytoplasmic extensions and invagina tions suggestive of niicropinacytosis. The cell membranes are continuous and show no pro toplasmic bridging or syncytial structures.

Electronmicroscopy provides some of the most persuasiveevidence that reef corals possess the necessary structural organization for transport of dissolved organic matter across the epidernial barrier. In all species so far examined the epidermis is shown to consist of tall columnar cells the free surface of which bears a single flagellum set into a shallow pit bordered by a circlet of nine to twelve @ microvilli about 2 long and 100 in diameter (Goreau and Philpot, 1956). The arrangement of the microvilli(Fig. 5) suggests they are a modified collar reminiscent of that of choanocytes. The surface membrane shows numerous and very variable cytoplasmic extensions and invaginations suggestive of micropina cytosis. Just beneath the surface are mitochondria and endoplasmic reticulum. Thus, not only are the epidermal cells shown to have a surface area many times greater than purely geometric estimates based on light microscopy would indi cate, but their ultrastructure is suggestive of a very dynamic cell boundary across TROPHIC CONDITIONS OF REEF CORALS 255

FIGURE 6. Tangential section through the epidermis of Colpopizyllia natans, stained for alkaline phosphatase an(l counterstained with eosin Y. The large lacunae represent mature mucus glands (muc gd). The phospliatase activity is confined to the supporting cells (epi sup c) here seen in cross section. Note that the enzyme has a reticular localization in what appear to be cell membranes. The small phosphatase negative vacuoles (muc) may be due to an early stage in the formation of mucus; magnification is X 5000. FIGURE 7. Tangential section through the epidermis of Colpo/'Izyllia natans stained with P. A. S. Compare with figure 6 which shows a serial section stained for phosphatases, and note that both this enzyme and the P. A. S. reactive substance have a similar localization confined to the supporting cells (el)i sup c). The mucus glands (muc gd) are negative; magnification is X 3600. 256 THOMAS F. GOREAU, NORA I. GOREAU AND C. M. YONGE which active transport takes place : the data from the amino acid uptake experi nients suggest that the net flux is from outside to inside. Histochemical studies (Goreau, 1956) have shown that highly active nonspecific alkaline phosphomonoesterases are present in the distal parts of the epidermal cells of corals (Fig. 6) . The precise function of these phosphatases is not clear : their extremely sharp and high pH maxima with peaks around pH 11.1 suggest that these enzymes do not act in vivo as simple hydrolases, but possibly as phospho transferases supplying energy for metabolic processes occurring in the outer sur face of the epidermis. We have not yet been able to establish on the ultrastructural level whether the alkaline phosphomonoesterase is associated with the microvilli. It is of considerable interest, however, that the localization of the enzyme within the coral epidermis is identical with that of a P. A. S. reactive non-metachromatic neutral mucopolysaccharide (Fig. 7) . A similar spatial association of alkaline phosphatase and neutral mucopolysaccharide was found by Moog and Wenger (1952)inabsorptiveandsecretoryorgansofseveralvertebrateandinvertebrate groups. In spite of their phyletic disparity, the epidermal cells of scleractinian corals and the absorptive epithelia of mammalian kidney and duodenum are re markably similar in general features of their ultrastructure, histochemistry and functions, having in common a large free surface area due to microvilli, high con centrations of alkaline phosphomonoesterases associated with neutral mucopoly saccharide at the free cell border, and being capable of active transport of dis solved organic substances against a concentration gradient. In view of these considerations, it is not unlikely that these epithelia also perform similar functions.

DIscussIoN After many years of controversy, much remains to be learnt about the nutrition of hermatypic corals. It is significant that, in distinction to ahermatypes, they exhibit a wide range in size and form of the polyps. This could well indicate a correspondingly wide range of specialization for dealing with food material ex tending from living animals to detritus and to particulate or dissolved material of animal origin. Corals such as Favia, Euphyllia or Mussa with large polyps can be observed to feed exclusively on animal prey, e.g., small fish and large zooplanktonic orga nisms or fragments of flesh (never vegetable matter), in precisely the same man ner as do ahermatypic corals such as Tubastrea or Balanophyllia and all Actiniaria. But, to the extent that these may be available, they may also absorb dissolved or colloidal matter through the epidermis by the mechanisms described above. Where polyps are smaller but still possess adequate tentacles, for instance, Porites or Pocillopora, and where ciliary currents beat toward instead of away from the mouth (Yonge, 1930a), a primary diet of smaller planktonic animals with par ticulate and/or dissolved organic matter may reasonably be postulated. In the extreme case of the agaricids (which are very common on reefs) such as Pavona, Psammocora or Agaricia with minute, and in some cases (e.g., Pachyseris) non existent, tentacles around very small mouths, particulate food must consist almost entirely of fine fragments of organic matter from the smallest zooplanktonic organisms downward. In such corals ciliary currents would appear to assist the boundary layer water TROPHIC CONDITIONS OF REEF CORALS 257 in conveying the finest material across the surface where the stimulus of animal matter in any form (down to amino acids) will cause mouths to open and mesenterial filaments to be extruded through them. These remarkably efficient organs for combined digestion and absorption of animal matter here take over the function of the tentacles. They extend out of the mouth or other openings in the tissue to seize and enwrap food particles which they may digest and absorb outside the coelenteron (Yonge, 1930a ; Abe, 1938) . In no scieractinian are the filaments reduced as they are in alcyonarians such as Xenia. While it is now abundantly established that material does pass from the zooxanthellae into the tissues of the host coelenterate—actiniarian, zoanthid or scleractinian (Goreau and Goreau, 1960 ; Muscatine, 1967, 1969 ; Von Holt and Von Holt, 1968; Trench, 1971a, 1971b, 1971c; Lewis and Smith, 1971)—the precise significance of this, in the context of the nutrition of the animal, still remains to be determined. Certainly the few species of temperate water actinians which harbor zooxanthellae appear in no way more efficient than the majority which do not. In the bivalve Tridacnidae there is an equally well established passage of soluble material into the blood stream from zooxanthellae (which are later digested in phagocytic blood cells) (Yonge, 1936 ; Goreau, Goreau and Yonge, 1966) . This material rapidly becomes incorporated into the byssus, crystalline style, periostracum and mucus indicating its possible use in the synthesis of these secretions in a manner similar to that described in the sacoglossan gastro pod, Tridachia (Trench, 1969 ; Trench, Greene and Bystrom, 1969). Maintained in darkness, some scleractinians can survive the eventual loss of the zooxanthellae, others cannot. This may not necessarily imply that the latter are suffering from starvation, it may equally be a consequence of the change in the internal environment. Normally the zooxanthellae automatically remove the waste products of metabolism, notably CO2 with sources of sulphur, nitrogen and phos phorus needed for protein synthesis. Not all hermatypic scleratinians may have the capacity for the efficient removal of these. Franzisket (1970) describes how, after exposure to darkness and consequent loss of zooxanthellae, the tissues of Porites atrophy but when exposed to light and reinfected with algae regeneration rapidly occurred. But it remains to be determined whether atrophy and subse quent recovery were the consequences of removal and then restoration of food supplied by the zooxanthellae or of inadequate metabolism when deprived of the algae which normally remove waste products. The major problems ahead in volve adequate evaluation of the precise energy needs of corals and the nature of available supplies—zooplankton and particulate or dissolved organic matter of animal origin—in coral reef seas. The spectacular success as reef builders of the hermatypic Scleractinia has tended to obscure the very small amount of living tissue actually present and so exaggerate the amount of food required.

Apart from the discussion, this paper was in rough draft at the time of Professor Thomas F. Goreau's death. Nora Goreau and Maurice Yonge wish to express their gratitude to Willard Hartman, David Barnes, Judith Lang and Rob ert Trench for criticism of the manuscript. We acknowledge with thanks the assistance of Peter Hunt who photographed our autoradiogram for Figure 4 and further photographic help from Thomas J. Goreau and E. A. Graham. This 258 THOMAS F. GOREAU, NORA I. GOREAU AND C. M. YONGE work was in part supported by the National Science Foundation, the U. S. Office of Naval Research and the Public Health Service of the U. S. Department of Health, Education and Welfare.

SUMMARY The assumption that reef corals are wholly autotrophic due to the presence of zooxanthellae is questioned. Reef corals lack the behavioral and structural spe cializations for an autotrophic existence comparable to that found in the xeniid octo corals and zoanthideans which appear to depend upon zooxanthellae for their food. The heterotrophic nutritional activities of reef corals, as observed both in the field and in the laboratory, include the following : ( 1) specialized carnivorous feeding, primarily on zooplankton, facilitated by ciliated currents and mucus, direct transfer of prey to the mouth by the tentacles, or extracoelenteric feeding by the mesenterial filaments ; (2) unspecialized detritus feeding, involving the use of a wide range of organic matter of animal and perhaps of bacterial origin ; (3) direct utilization of dissolved or colloidal organic matter as suggested by the uptake of amino acids by the epidermis and by the ultrastructural, histochemical and physio logical features of the free cell border. Water circulating within the reef, the boundary layer water, is in a continuous and dynamic exchange with the trophic structure of the reef, recycling nutrients with the benthos and making the suspended particulate matter a possible food source for corals.

LITERATURE CITED

ABE, N., 1938. Feeding behaviour and the nematocyst of Fungia and 15 other species of corals. Palao Trop. Biol. Sta. Stud., 1: 469—521. BAKUS, G. J., 1967. Some relationships of fishes to benthic organisms on coral reefs. Nature, 210: 280—284. BAKUS, G. J., 1969. Energetics and feeding in shallow marine waters. mt. Rev. Gen. Ezp. Zool., 4: 275—369. BARDACH, J. E., 1961. Transport of calcareous fragments by reef fishes. Science, 133: 98—99. BAYLOR, E. R., AND W. H. SUTCLIFFE, 1963. Dissolved organic matter in seawater as a source of particulate food. Limnol. Oceanogr., 8: 369—371. DUERDEN, J. E., 1902. West Indian Madreporarian polyps. Mem. Nat. Acad. Sci., 8: 399— 597. DUERDEN, J. E., 1906. The role of mucus in corals. Quart. J. Microscop. Sci., 49: 591—614. FLEMING, M. W., 1954. Plankton of Northern Marshall Islands U. S. Geol. Surv. Prof. Pap., 260(F): 301—314. FRANZISKET, L., 1969a. Riff Korallen Können autotroph Leben. Naturwissenshaf ten, 56(3): 144. FRANZISKET, L., 1969b. The ratio of photosynthesis to respiration of reef building corals during a 24 hour period. Forma et Functio, 1: 153-158. FRANZISKET, L., 1970. The atrophy of hermatypic reef corals maintained in darkness and their subsequent regeneration in light. mt. Rev. Hydrobiol., 55: 1—12. GOHAR, H. A. F., 1940. Studies on the Xeniidae of the Red Sea. Publ. Mar. Biol. Sta. Ghardaqa, Red Sea, Egypt, No. 2: 25—118. GOHAR, H. A. F., 1948. A description and some biological studies of a new alcyonarian speciesClavulariahamra Gohar. Pubi.Mar. Biol.Sta.Ghardaqa,Red Sea,Egypt, No. 6: 3—33. GORDON,M. S., AND H. M. KELLEY, 1962. Primary productivity of an Hawaiian coral reef: a critique of flow respirometry in turbulent waters. Ecology, 43: 473—480. TROPHIC CONDITIONS OF REEF CORALS 259

GOREAU, T. F., 1953. Phosphomonoesterases in reef building corals. Proc. Nat. Acad. Sci., 39: 1291—1295. GOREAU, T. F., 1956. Histochemistry of mucopolysaccharide-like substances and alkaline phos phatase in Madreporaria. Nature, 177 : 1029—1030. GOREAU, T. F., AND N. I. GOREAU, 1960. Distribution of labelled carbon in reef-building corals with and without zooxanthellae. Science, 131 : 668—669. GOREAU, T. F., AND E. A. GRAHAM, 1967. A new Halimeda from Jamaica. Bull. Mar. Sci., 17: 432—441. GOREAU, T. F., AND W. D. HARTMAN, 1963. Boring sponges as controlling factors in the formation and maintenance of coral reefs. Pages 25—54in R. F. Sognnaes, Ed., Mechanisms of Hard Tissue Destruction. A;ner. Assoc. Advan. Sci. Publ., No. 75, Washington, D. C. GOREAU, T. F., AND D. E. PHILPOT, 1956. Electronmicrographic study of flagellated epithelia in Madreporarian corals. Exp. Cell Res., 10 : 552—555. GOREAU, T. F., AND J. W. WELLS, 1967. The shallow water Scleractinia of Jamaica : revised list of species and their vertical distribution ranges. Bull. Mar. Sci., 17 : 442—453. GOREAU, T. F., N. I. GOREAU AND C. M. YONGE, 1966. Evidence for a soluble algal factor produced by the zooxanthellae of Tridacna elongata. Abstract. mt. Conf. Trap. Oceanography, Miami, 1965. [Available from Department of Zoology, University of Edinburgh, Scotland.] GOREAU, T. F., N. I. GOREAU, C. M. YONGE AND Y. NEUMANN, 1970. On feeding and nutrition in Fungiacava eilatensis Soot-Ryen (Bivalvia, Mytilidae), a commensal living in fungiid corals. J. Zoo!. London, 160 : 159—172. GOREAU, T. F., V. TORRES, L. MAS AND E. RAMOS, 1960. On community structure, standing crop and oxygen balance of the Lagoon at Cayo Turrumote, Puerto Rico. (Abstract). Assoc. Island Mar. Labs. 3rd Meeting, Jamaica. HARTMAN, W. D., AND T. F. GOREAU, 1970. Jamaican coralline sponges : their morphology, ecology and fossil relatives. Sy;np. Zoo!. Soc. London, 25 : 205—243. HELLEBUST, J. A., 1965. Excretion of some organic compounds by marine phytoplankton. Limnol. Oceanogr., 10: 192—206. JACKSON, J. B. C., T. F. GOREAU AND W. D. HARTMAN, 1970. Recent brachiopod sclero sponge communities and their palaeoecological implications. (In press) JOHANNES, R. E., AND S. L. COLES, 1969. The role of zooplankton in the nutrition of scleractiniancorals.[Symp. Corals,Coral Reefs.Jan. 1969]. Mar. Biol.Assoc. india, Mandapani Camp, 1969: 8. KOHN, A. J., AND P. HELFRICH, 1957. Primary organic productivity of a Hawaiian coral reef. Limno!. Oceanogr., 2: 241—251. LANG, J. C., 1969. New characters for coral taxonomy. (Abstract) Assoc. Island Mar. Labs. 8th Meeting, Jamaica. LANG, J. C., 1970. Inter-specific aggression within the scleractinian reef corals. Ph.D. thesis, Yale University. LEWIS, D. H., AND D. C. SMITH, 1971. The autotrophic nutrition of symbiotic marine co elenterateswith specialreferenceto hermatypiccorals.I. Movement of photo synthetic products between the symbionts. Proc. Roy Soc. London Series B, 178: 111—129. @V1ARISCAL,R. N., AND H. M. LENHOFF, 1968. The chemical control of feeding behaviour in Cyphastrea ocellina and some other Hawaiian corals. J. Ezp. Biol., 49: 689—699. MARSHALL, N., 1965. Detritus over the reef and its potential contribution to adjacent waters of Eniwetok Atoll. Ecology, 46: 343—344. MATTHAI, G., 1918. On reactions to stimuli in corals. Phil. Soc. Proc. Cambridge, 19: 164- 166. MOOG, F., AND E. L. WENGER, 1952. The occurrence of a neutral mucopolysaccharide at sites of high alkaline phosphatase activity. Amer. J. Anat., 90: 339—378. MUSUATINE, L., 1967. Glycerol excretion by symbiotic algae from corals and Tridae,,a and its control by the host. Science, 156: 516—519. MISCATINE, L., AND E. CERNICHIARI, 1969. Assimilation of photosynthetic products of zoo xanthellae by a reef coral. [Hal. Bull., 137 (3) : 506—523. 260 THOMAS F. GOREAU, NORA I. GOREAU AND C. M. YONGE

ODUM, H. T., AND E. P. ODUM, 1955. Trophic structure and productivity of a windward coral reef community on Eniwetok Atoll. Ecol. Monogr., 25 : 291—320. RILEY, G. A., 1963. Organic aggregates in seawater and the dynamics of their formation and utilization. Limnol. Oceanogr., 8 : 372—381. SARGENT, M. C., AND T. S. AUSTIN, 1949. Organic productivity of an atoll. Trans. Amer. Geophys. Union, 30 : 245—249. SARGENT, M. C., AND T. S. AUSTIN, 1954. Biologic economy of coral reefs. U. S. Geol. Sur@'. Prof. Pap., 260(E) : 293—300 STEPHENS, G. C., 1962. Uptake of organic matter by aquatic invertebrates. I. Uptake of glucose by the solitary coral Fungia scutaria. Biol. Bull., 123: 648-659. STODDART, D. R., 1969. Ecology and morphology of recent coral reefs. Biol. Rev., 44 : 433— 498. TRENCH, R K., 1969. Chloroplasts as functional endosymbionts in the mollusc Trida€hia ens pata (Bergh), (Opisthobranchia, Sacoglossa). Nature, 222 : 1071—1072. TmaNcH, R. K., 1971a. The physiology and biochemistry of zooxanthehlae symbiotic with marine coelenterates. I. The assimilation of photosynthetic products of zooxantheliae by two marine coelenterates. Proc. Roy. Soc. London Series B, 177: 225—235. TRENCH, R. K., l97lb. The physiology and biochemistry of zooxanthellae symbiotic with marine coelenterates. II. Liberation of fixed ‘¿4Cby zooxanthellae in vitro. Proc. Roy. Soc. London Series B, 177: 237—250. TRENCH, R. K., 1971c. The physiology and biochemistry of zooxanthellae symbiotic with marine coehenterates. III. The effect of homogenates of host tissues on the excretion of photosynthetic products i,z vitro by zooxanthellae from two marine coelentrates. Proc. Roy. Soc. Lo;z4on Series B, 177: 251—264. TRENCH, R. K., R. W. GREENE AND B. G. BYSTROM, 1969. Chloroplasts as functional organelles in animal cells. J. Cell. Biol., 42: 404—417. VON HOLT, C.,1968. Uptake of glycineand releaseof nucleosidepolyphosphatesby zoo xanthelhae. Comp. Biocizeni. Physiol., 26: 1071—1079. VoN HOLT, C., AND M. VON H0LT, 1968a. Transfer of photosynthetic products from zooxanthellae to coehenterate hosts. Comp. Biochem. Physiol., 24: 73—81. VON HOLT,C.,AND M. VON HOLT, 1968b.The secretionoforganiccompoundsby zooxanthel lae isolated from various types of Zoanthus. Cornp. Biochern. Physiol., 24: 83—92. YONGE, C. M., 1930a. Studies on the physiology of corals. I. Feeding mechanisms and food. Sci. Rep. Great Barrier Reef Exped., 1: 13-57. YONGE, C. M., 1930b. Studies on the physiology of corals. III. Assimilation and excretion. Sci. Rep. Great Barrier Reef Exped., 1: 83—91. YONGE, C. M., 1936. Mode of life, feeding, digestion and symbiosis with zooxanthellae in the Tridacnidae. Sci. Rep. Great Barrier Reef Exped (1928—29),1:283—321. YONGE, C. M., 1940. The biology of reef building corals. Sci. Rep. Great. Barrier Reef Exped., 1: 353—391. YONGE, C. M., AND A. G. NICHOLS, 1930. Studies on the physiology of corals. II. Digestive enzymes. Sci. Rep. Great Barrier Reef Exped., 1: 59—81. YONGE, C. M., AND A. G. NICHOLLS, 1931. Studies on the physiology of corals. IV. The structure, distribution and physiology of the zooxanthellae. Sci. Rep. Great Barrier Reef Exped., 1: 135—176.